Rice Disease Classification using Deep Learning

Abstract

Rice is one of the essential foods for most of the world’s population, and its cultivation is done in large areas of Pakistan. It satisfies the food demand and also plays a vital role in the economy of Pakistan. But rice leaf diseases can affect its production. Rice leaf diseases can be categorized into various diseases like brown spots, bacterial blight, tungro, and leaf blasts. Early detection of rice leaf diseases is critical for the effective management and prevention of crop losses. Effective treatment of these diseases can be done after their detection. Detection of such diseases can be performed using a deep learning approach. The small dataset allows transfer learning models and enhanced Xception models used to detect diseases effectively. The Xception model has achieved 93.92% test accuracy when the epoch value is set to 10. Model evaluation has also been done using accuracy, recall, precision, and confusion matrix. Deep learning approaches for the detection of rice leaf diseases in Pakistan can have a positive impact on rice production and the country's economy.